Democratizing ML for speech (Practical AI #164)
Changelog Master Feed
00:00
The Snake Is Eating Its Own
Machine learning is very good at handling kind of rough inputs, and so it seemed like a pretty good hypothesis there. The ways that a machine in the loop for labelling go wrong are going to be probably different than a human. I guarantee you, like any engine er, we tried a few things, and the first things did not work. We got in the ball park of libera speech. Which were like, that's good enough. But i will tell you, we did not start out in the ball Park of liberaspeech. There is all sorts of problems, you know. You have like, mismatched transcripts where things are just, but o, you find them down one
Transcript
Play full episode
